Review of Finance (2010) 14: 235–262

doi: 10.1093/rof/rfq005
Advance Access publication: 6 March 2010

Determinants of Sovereign Risk: Macroeconomic

Fundamentals and the Pricing of Sovereign Debt∗

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

JENS HILSCHER1 and YVES NOSBUSCH2
1
Brandeis University; 2 London School of Economics

Abstract. This paper investigates the effects of macroeconomic fundamentals on emerging market
sovereign credit spreads. We find that the volatility of terms of trade in particular has a statistically
and economically significant effect on spreads. This is robust to instrumenting terms of trade with
a country-specific commodity price index. Our measures of country fundamentals have substantial
explanatory power, even controlling for global factors and credit ratings. We also estimate default
probabilities in a hazard model and find that model implied spreads capture a significant part of the
variation in observed spreads out-of-sample. The fit is better for lower credit quality borrowers.

JEL Classification: F34, G12, G13, G15

1. Introduction

There is tremendous variation in the interest rates emerging market governments

pay on their external debt. This is true both across countries and over time. A
common measure of a country’s borrowing cost in international capital markets is
its yield spread, which is defined as the difference between the interest rate the
government pays on its external U.S. dollar denominated debt and the rate offered
by the U.S. Treasury on debt of comparable maturity.

∗
We are grateful to John Campbell, N. Gregory Mankiw, Kenneth Rogoff, and Jeremy Stein for
their advice and suggestions. We thank an anonymous referee, Alberto Alesina, Silvia Ardagna,
Gregory Connor, Darrell Duffie, Nicola Fuchs-Schündeln, Amar Gande, David Hauner, Dirk Jenter,
David Laibson, David Lesmond, Emi Nakamura, Ricardo Reis, Andrei Shleifer, Monica Singhal,
Jón Steinsson, Federico Sturzenegger, Suresh Sundaresan, Glen Taksler, Sam Thompson, Don van
Deventer, and seminar participants at Harvard University, the Federal Reserve Board of Governors,
Warwick Business School, Oxford Saı̈d Business School, the Bank for International Settlements,
Macalester, UC Irvine, Brandeis University, Barclays Global Investors, Oak Hill Platinum Partners,
the 2006 MMF and FMA conferences, the Swiss National Bank, the 2007 Moody’s and Copenhagen
Business School Credit Risk Conference, the 2007 EEA and EFA meetings, the Bank of England,
the 2008 ESRC/WEF Warwick Conference on Incentives and Governance in Global Finance, NUI
Maynooth, Exeter University, and Bristol University for helpful comments and discussions. We also
thank Carmen Reinhart, the IMF, J.P. Morgan, and the World Bank for providing us with data, and
Pavel Bandarchuk and Manjola Tase for research assistance.


C The Authors 2010. Published by Oxford University Press [on behalf of the European Finance Association].
All rights reserved. For Permissions, please email: [email protected]
236 JENS HILSCHER AND YVES NOSBUSCH

In this paper we consider to what extent macroeconomic fundamentals can ex-

plain variation in sovereign yield spreads. We focus in particular on the explanatory
power of the volatility of fundamentals. From a theoretical perspective, the volatil-
ity of fundamentals should matter for the pricing of defaultable debt, just as the
level of these fundamentals does. All else equal, a country with more volatile fun-
damentals is more likely to experience a severe weakening of fundamentals which
may force it into default. This risk should be reflected in a higher yield spread on
its bonds.1 While this idea is well understood in theory, it has been largely ignored

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

by the empirical literature on sovereign debt, which has tended to focus on level
variables.2 In contrast, we include both the level and the volatility of macroeconomic
fundamentals in our empirical analysis.
There are two main parts to the paper. In the first part, we analyze the effect
of country-specific fundamentals and global factors on sovereign debt prices for a
set of 31 emerging market countries from 1994 to 2007. We measure yield spreads
on external U.S. dollar denominated debt using data from J.P. Morgan’s Emerging
Market Bond Index (EMBI). In the second part, we examine the effect of these
macroeconomic variables on the probability of default, using data from 1970 to
2007. We also relate out-of-sample fitted default probabilities to observed spreads.
We find that macroeconomic fundamentals have statistically and economically
significant effects on spreads. We focus in particular on terms of trade, which
measure the price of a country’s exports relative to its imports. As noted by Bulow
and Rogoff (1989), changes in a country’s terms of trade affect its ability to generate
dollar revenue from exports and therefore its ability to make payments on its external
dollar denominated debt. The volatility of terms of trade also matters for the overall
economy: terms of trade volatility is important for explaining output variability at
business cycle frequencies (Mendoza, 1995) and has negative effects on long run
growth (Mendoza, 1997). We find that spreads indeed tend to be higher for countries
that have recently experienced adverse terms of trade shocks, while countries that
have seen their terms of trade improve tend to have lower spreads. We also find
that the volatility of terms of trade has a highly significant effect on spreads, both
statistically and economically.
One concern is that the terms of trade could be partly endogenous. In order
to address this issue, we construct a country-specific commodity price index. Our
index includes only major commodity prices traded in world markets and we weight
these prices by country-specific export shares. We believe that this index is plausibly
1
In particular, this intuition holds in Merton’s (1974) seminal model of risky corporate debt.
2
Exceptions are Edwards (1984), who includes variability of reserves and finds that it is insignificant,
Westphalen (2001), who finds some limited effect of changes in local stock market volatility on
changes in short term debt prices, and Garcia and Rigobon (2005), who find that risk-based measures
of debt sustainability are closely related to spreads in the case of Brazil. In the corporate bond context,
Campbell and Taksler (2003) find a strong empirical link between equity volatility and yield spreads.
DETERMINANTS OF SOVEREIGN RISK 237

exogenous (Chen and Rogoff, 2003) and use it as an instrumental variable for terms
of trade. The instrumentation does not lead to a significant change in the coefficients
on our main variables or the overall regression fit.
We also examine the relative importance of country-specific and global factors.
A number of papers have emphasized the importance of global factors for emerging
markets: Calvo et al. (1993), and more recently Calvo (2002), Herrera and Perry
(2002), Grandes (2003), Diaz Weigel and Gemmill (2006), Garcı́a-Herrero and
Ortiz (2006), Longstaff et al. (2007), and González-Rozada and Levy Yeyati (2008).

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

In order to control for global factors, we include the implied volatility of the S&P500
index (VIX), the U.S. default yield spread, the 10-year U.S. Treasury yield, and
the difference between the 3-month Libor rate and the 3-month Treasury rate
(TED spread). We find that global factors are indeed important. In particular, the
coefficient on the VIX index is positive and significant, which is consistent with Pan
and Singleton (2008) who find that the VIX is statistically significant in explaining
credit default swap (CDS) spreads of Mexico, Turkey, and Korea. Nevertheless, we
find that country-specific fundamentals have substantial explanatory power, even
after controlling for global factors.
Our measures of country-specific fundamentals also add explanatory power
relative to credit ratings, which are often used as a summary measure of credit
quality (e.g. Cantor and Packer, 1996; Kamin and von Kleist, 1999; González-
Rozada and Levy Yeyati, 2008; Powell and Martinez, 2008). Including our country-
specific variables leads to a substantial increase in adjusted R2 compared to a
regression that only includes global variables and credit ratings and the coefficients
on our country-specific variables remain significant when we include ratings.
Finally, we note that, even after accounting for macroeconomic fundamentals,
spreads are higher for countries that have recently emerged from default. This is
consistent with Reinhart et al. (2003), who find that a key predictor of future default
is a country’s history of default. Some countries seem to behave like “serial default-
ers” in a way that cannot fully be accounted for by macroeconomic fundamentals.
In the second part of the paper, we estimate default probabilities and use fitted
probabilities to price debt. This allows us to directly determine the extent to which
fundamentals affect spreads through their effect on default probabilities. Using
data on country defaults, we estimate the conditional probability of default over
the next period in a reduced form logit model following the approach proposed
by Shumway (2001) in the context of predicting corporate bankruptcy. We find
that the volatility of terms of trade is an important predictor of default, in line
with its effect on spreads. We then use estimated default probabilities to calculate
out-of-sample predicted spreads and find that predicted spreads can account for a
substantial share of the variation in observed spreads. Our methododolgy in this part
of the paper is related in spirit to the approach of Garcia and Rigobon (2005). These
authors propose a risk-based model of debt sustainability that takes into account
238 JENS HILSCHER AND YVES NOSBUSCH

the stochastic nature of a country’s debt accumulation equation. This allows them
to infer the probability that debt to GDP exceeds a given threshold. They show that
these model implied probabilities are closely related to EMBI spreads in the case
of Brazil. Calvo and Mendoza (1996) present related evidence in the context of the
Mexican crisis of 1994–95.3
When we group bond spread observations into quintiles by their estimated de-
fault probability we find that predicted spreads are smaller than actual spreads
for all the groups,4 but that the proportion of the spread explained by the default

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

prediction model increases significantly with the probability of default. Huang and
Huang (2003) find a similar pattern when calibrating structural form models to
U.S. corporate bond prices.
This paper adds to the large literature on the empirical determinants of sovereign
yield spreads. The literature varies widely in choice of variables and methodology.
Some papers concentrate on reduced form regressions of spreads on explanatory
variables. Examples of this line of research include Edwards (1986), Eichengreen
and Mody (1998), Min (1998), Beck (2001), and Ferrucci (2003), among others.
These authors explore a large set of macroeconomic variables to explain spreads.
As already mentioned, we pay special attention to measures of volatility which are
largely absent from this literature.
Duffie et al. (2003) develop a flexible reduced form model of sovereign
yield spreads. They estimate their model using weekly data on Russian dollar-
denominated debt and U.S. swap yields between 1994 and 1998. Pan and Singleton
(2008) explore the term structure of CDS spreads for Mexico, Turkey, and Korea
from 2001 to 2006 and consider the risk-neutral credit event intensities and loss
rates that best describe the CDS data. While these studies relate spreads implied
by a reduced from pricing model to political factors, foreign currency reserves, oil
prices, and VIX, macroeconomic fundamentals do not directly enter the estimation
of the pricing model. In contrast, our focus is to explore directly the extent to
which variation in macroeconomic fundamentals determines spreads and default
probabilities across countries and over time.
There is also a large literature on predicting banking and currency crises and
sovereign defaults using macroeconomic fundamentals.5 Again, the focus of these

3
In the literature on corporate default risk, the link between default probabilities and spreads has
been explored by a number of papers, e.g. Anderson and Sundaresan (1996), Huang and Huang
(2003), and Berndt et al. (2005).
4
This is not surprising given that we only model the default component of the spread. We discuss
other spread components in Section 4.
5
Early contributions to this literature include Hajivassiliou (1987, 1994). Berg et al. (2000),
Kaminsky and Reinhart (1999), and Goldstein et al. (2000) have concentrated on constructing what
they refer to as “early warning systems.” Berg et al. (2005) provide an overview of this line of
research.
DETERMINANTS OF SOVEREIGN RISK 239

studies is on level variables, rather than volatility. There are some exceptions: Catão
and Sutton (2002) and Catão and Kapur (2006) predict sovereign default in a hazard
model using a similar setup to the one used in the second part of this paper. They
identify volatility of terms of trade as an important predictor of default.
The remainder of the paper is organized as follows. Section 2 describes the
data. Section 3 describes the results from using fundamentals to explain spreads
in a regression framework. In Section 4 we turn to default prediction and compare
out-of-sample predicted spreads to observed spreads. Section 5 concludes.

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

2. Choice of Variables and Data Description

Our empirical investigation concentrates on explaining variation in sovereign

external debt prices across countries and over time. We restrict ourselves to
dollar-denominated debt instruments issued or guaranteed by emerging mar-
ket governments.6 Our measure of the yield spread over U.S. Treasuries is J.P.
Morgan’s Emerging Markets Bond Index Global (EMBI Global). This series con-
sists of daily data starting in January of 1994 and covers a sample of 32 emerging
market countries.7 It includes Brady bonds, loans, and Eurobonds with an average
maturity of 12 years and verifiable prices. We discuss inclusion requirements in
more detail in the appendix.
Table I lists the set of countries with available EMBI spread data and groups
countries into regions. The largest regional group is made up of countries in Latin
America with 12 countries. The data set also includes 5 countries in Africa, 6 in
Eastern Europe and a total of 9 in the Middle East and Asia. The spread data begin
in 1994, but not all countries have data going back that far. We report the first year
of available spread data for each country. By 1996, nearly half of the eventual set
of countries have available data and since 1998 J.P. Morgan has been constructing
series for close to two thirds of the countries.
There is tremendous variation in spreads both across countries and over time.
Table I reports median EMBI spreads from 1998 to 2007 ranging from 62 basis
points (Hungary) to 2666 basis points (Côte d’Ivoire).8 Figure 1 plots the weighted
average EMBI spread.
We begin our investigation of spread determinants by considering some sugges-
tive global evidence that terms of trade are an important determinant of sovereign
6
In contrast, the yield on local currency debt tends to be driven mainly by local inflation risk.
7
The final spread regression sample contains only 31 countries; due to missing macroeconomic
data Nigeria is not included. However, Nigeria is included in the longer default prediction sample
(Section 4).
8
The large variation in spreads is driven partly by observations of countries in default, e.g. Argentina,
Nigeria, and Côte d’Ivoire. Exluding these, the highest two median spreads are equal to 1305 and
565 basis points.
Table I. Summary statistics by country

240
The table presents statistics by country. EMBI refers to J.P. Morgan’s Emerging Market Bond Index Global (EMBI). The spread is measured in basis
points over U.S. Treasuries. We report the first year for which EMBI data is available and the median end-of-year spread between 1998 and 2007.
Volatility of terms of trade is the median standard deviation of the percent change in terms of trade calculated using annual data and a rolling ten year
backward-looking window. Number of defaults is the number of times a country has gone into default over the next year. Debt/GDP is the median
level of external debt divided by GDP. These three colums are for the sample period 1970 to 2007. Export groups are for 2000–2001 (UNCTAD).
Start of Volatility
Country EMBI EMBI of terms Number of Top two exports (SITC group) and their percentage
Country code series spread of trade defaults Debt/GDP of total exports

Latin America Argentina ARG 1994 707 9.2 2 42.4 Crude petroleum (10%), feeding stuff for animals (10%)
Brazil BRA 1994 565 9.5 1 33.1 Aircraft (6%), iron ore and concentrates (5%)
Chile CHI 1999 143 9.6 1 46.4 Copper (27%), base metals ores (13%)
Colombia COL 1997 436 10.0 0 32.7 Crude petroleum (26%), coffee and substitutes (8%)
Dominican Rep. DMR 2001 447 0.9 1 29.4 Men’s outwear non-knit (17%), under garments knitted (13%)
Ecuador EQU 1995 824 12.4 3 60.9 Crude petroleum (41%), fruit, nuts, fresh, dried (18%)
El Salvador ESD 2002 238 13.6 0 30.5 Coffee and substitutes (16%), paper and paperboard, cut (6%)
Mexico MEX 1994 310 7.8 1 31.7 Passenger motor vehicles, exc. bus (10%), crude petroleum (8%)
Panama PAN 1996 346 2.3 1 77.6 Fish, fresh, chilled, frozen (20%), fruit, nuts, fresh, dried (20%)
Peru PER 1997 435 14.5 4 52.3 Gold, non-monetary (17%), feeding stuff for animals (13%)
Uruguay URU 2001 308 8.4 4 34.7 Meat, fresh, chilled, frozen (16%), leather (10%)
Venezuela VEN 1994 601 21.8 3 40.3 Crude petroleum (59%), petroleum products, refined (25%)
Africa Côte d’Ivoire CDI 1998 2666 19.7 2 108.0 Cocoa (28%), petroleum products, refined (18%)
Morocco MOR 1997 429 7.3 2 55.6 Women’s outwear non-knit (11%), men’s outwear non-knit (8%)

JENS HILSCHER AND YVES NOSBUSCH

Nigeria NIG 1994 1061 26.3 1 52.3 Crude petroleum (99.6%)
South Africa SAF 1994 236 5.4 3 16.9 Pearl, prec., semi-prec. stones (13%), special transactions (12%)
Tunisia TUN 2002 118 4.8 0 56.2 Men’s outwear non-knit (17%), women’s outwear non-knit (10%)
Eastern Europe Bulgaria BUL 1994 285 5.3 1 66.0 Petroleum products, refined (11%), copper (7%)
Croatia CRO 1996 353 1.5 1 59.3 Ships, boats etc. (15%), petroleum products, refined (8%)
Hungary HUN 1999 62 2.9 0 58.7 Int. combust. piston engines (9%), automatic data proc. eq. (7%)
Poland PLD 1994 165 3.9 1 39.0 Furniture and parts thereof (7%), ships, boats etc. (4%)
Russia RUS 1997 450 13.1 2 41.6 Crude petroleum (24%), gas, natural and manufactured (17%)
Ukraine UKR 2000 290 12.1 1 36.9 Iron, steel primary forms (13%), iron, steel shapes etc. (8%)
Southeast Asia China CHN 1994 98 2.6 0 12.8 Telecom equip., parts, acces. (5%), automatic data proc. eq. (5%)
Malaysia MAL 1996 178 7.0 0 38.9 Transistors, valves etc. (19%), office, adp. mach. parts (12%)
Philippines PHI 1997 435 7.6 1 62.4 Transistors, valves etc. (42%), automatic data proc. eq. (13%)
South Korea SKO 1994 168 3.6 0 34.6 Transistors, valves etc. (12%), passgr. motor vehicl. exc. bus (7%)
Thailand THA 1997 128 7.6 0 34.1 Office, adp. mach. parts (9%), transistors, valves etc. (8%)
Middle East & Egypt EGY 2001 131 9.6 1 40.5 Petroleum products, refined (39%), crude petroleum (10%)
South Asia Lebanon LEB 1998 369 3.8 0 28.1 Gold, silver ware, jewelry (9%), gold, non-monetary (7%)
Pakistan PAK 2001 339 11.7 1 43.9 Textile articles (16%), textile yarn (12%)
Turkey TUR 1996 379 5.0 2 34.8 Outer garments knit non-elastic (6%), under garments knitted (6%)

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

DETERMINANTS OF SOVEREIGN RISK 241

1500

2.7
1300
EMBI
EMBI spread (basis points over U.S. Treasuries)

Commodity_Index

1100

CRB commodity index

900

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

2.5
700

500

300

100 2.3
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
date

Figure 1. EMBI spreads and Commodity Prices

This figure plots EMBI spread (from J.P. Morgan, monthly spread over U.S. Treasuries), weighted by
country external debt, and the CRB commodity price index (in logs).

yield spreads. Since many emerging markets are important commodity exporters,
we expect an improvement in the terms of trade and a resulting decline in spreads
when commodity prices increase. We plot commodity prices as measured by the
Commodity Research Bureau (CRB) commodity index in Figure 1 and find that
commodity prices and spreads in fact exhibit a strong negative correlation of −0.66,
which is apparent from the graph.
A natural interpretation is that when commodity prices are high, commodity
exporters are more likely to repay their external debt, which reduces the yield
spread they face in international capital markets. For example, the substantial
increase in commodity prices from 2002 to 2007 was accompanied by significant
spread narrowing. However, it is important to note that we cannot readily generalize
from this evidence since countries have very different baskets of exports, the top
two of which we report in Table I. For example, Venezuela exports mainly oil, while
Chile exports copper and other metals. In order to capture these differences we
use more detailed country-specific terms of trade data in our subsequent empirical
investigation.

2.1 COUNTRY VARIABLES

A country’s ability to pay its external debt affects its probability of default and,
therefore, the spread it has to pay in international capital markets. In order to analyze
spread variation across countries and over time we gather annual cross-country data
242 JENS HILSCHER AND YVES NOSBUSCH

on macroeconomic fundamentals from a variety of sources. We provide details in

the appendix.
In our choice of country variables we focus on measures of fundamentals that
are likely to be related to the risk of default. We would expect a country’s terms
of trade, the price of the country’s exports relative to its imports, to affect its
ability to generate dollar revenue and repay its external debt. The intuition for
why this measure is relevant for the pricing of debt is apparent when considering
an oil-exporting country. The country generates dollar revenue by exporting oil

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

and spends dollars on imports. If the price of oil rises, the country is in a better
position to generate export revenue and repay its dollar-denominated liabilities.
Since terms of trade is a relative price series, we measure the country’s terms of
trade by constructing the percentage change in the terms of trade over the past five
years (Change in terms of trade). A positive number means that a country’s exports
have become more expensive relative to its imports.
Bondholders, however, do not only care about recent changes in the terms of
trade but also about the risk of a large adverse shock in the future, say a large
decline in the price of oil in our example. As a result, we expect the volatility of
terms of trade to be positively correlated with spreads. We calculate the standard
deviation of the percentage change in terms of trade (Volatility of terms of trade)
over the previous ten years.9 We graph terms of trade volatility against EMBI
spreads for our cross-section of countries in Figure 2, which indeed suggests a
positive relationship.10
We also include a measure of a country’s recent default history. This variable is
motivated by the work of Reinhart et al. (2003), who argue that history of default
is an important predictor of future default. These authors have constructed annual
default indicators for different debt categories for a large sample of countries going
back to the early nineteenth century. We use their series for default on total debt,
which includes foreign currency bonds and foreign currency bank debt. We count
the years since the country’s last year in default (Years since last default). We cap
the variable at 10 and set it equal to 11 if a country has never defaulted.11
We summarize default activity for the set of countries in Table I. There are a
total of 40 sovereign default episodes since 1970. A number of countries defaulted

9
For the few cases where data on terms of trade start later in the sample, we reduce the window size
to a minimum of six years.
10
There is considerable variation in terms of trade volatility across countries. High terms of trade
volatility is to a large extent driven by oil prices. Petroleum products are among the top two export
categories for the three countries with the highest terms of trade volatilities (Table I). However, our
results are robust to excluding countries with more than 20% of their exports concentrated in oil.
11
If unrestricted, this measure could become very large for countries that continue to stay out
of default. We make these adjustments because we expect additional years out of default to have
incrementally less importance.
DETERMINANTS OF SOVEREIGN RISK 243

CDI
1250

1000
EMBI spread (basis points)

EQU

VEN
750

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

BRA
ARG

BUL
500
URU PAK
LEBTUR
COL UKR
PHIMOR
PER
MEX
PAN
DMR RUS
250 ESD
SAFSKO
CRO
MAL PLD
EGY
TUN THA CHI
CHN
HUN

5 10 15 20 25
Volatility of terms of trade (%)

Figure 2. Volatility of terms of trade and EMBI spread

This graph plots mean EMBI spread (from J.P. Morgan, end-of-year spread over U.S. Treasuries)
against mean country volatility of terms of trade.

on multiple occasions – Peru and Uruguay each have four defaults – while nine
countries did not go into default during this period.

2.2 TERMS OF TRADE AND COMMODITY PRICES

While many of the countries in our data set are large commodity exporters whose
export prices are determined in world markets, not all of the countries’ exports
are concentrated in commodities. It is therefore possible that our measure of terms
of trade may be influenced by local factors, raising the concern of endogeneity.
In order to address this, we construct country-specific commodity export price
indices which we use to instrument for terms of trade in our empirical investigation
in Section 3.
We gather data on export quantities from COMTRADE and export prices from
Global Financial Data. We then construct an export-weighted price index for each
country using time-varying weights. The commodities included in the basket are
those for which we are able to match export price and quantity data. Specifically, we
include the following commodities (ordered by the mean share in country exports)
in the price indices: oil, coffee, textiles, copper, cotton, cocoa, meat, bananas,
leather, gold, gas, and silver (SITC codes listed in the appendix). We use these price
244 JENS HILSCHER AND YVES NOSBUSCH

indices to construct the change in export prices and the volatility of export price
changes analogous to the change in terms of trade and the volatility of terms of
trade discussed above. In the next section we use the commodity price index to
instrument for terms of trade. Before doing so we check that price index changes
are associated with terms of trade changes. We regress the terms of trade measures
on the price index measures and find that, for both measures, the coefficients are
statistically significant at the 1% level and the R 2 are close to 0.5.

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

2.3 GLOBAL AND CONTROL VARIABLES

In order to control for global factors such as changes in aggregate risk aversion,
world interest rates, and liquidity, we add four time series variables. We include
the VIX index and the U.S. default yield spread, defined as the spread between
corporate bonds with a Moody’s rating of Baa and Aaa. As a proxy for the world
interest rate we include the 10-year U.S. Treasury rate and to capture changes in
aggregate liquidity we include the TED spread.
We also examine two additional country-specific variables: external debt to GDP
(Debt/GDP) and the ratio of reserves (including gold) to GDP (Reserves/GDP). A
number of papers, particularly in the early literature on this topic (e.g. Edwards,
1986), have found a significant positive coefficient on the ratio of debt to GDP in
spread regressions. There are two main concerns with these variables. First, they
are both potentially endogenous. The dynamics of debt in particular are likely to
be endogenous and non-linear (e.g. Favero and Giavazzi, 2002; 2005). A country’s
GDP may also be affected by its spread (e.g. Uribe and Yue, 2006). Second, neither
variable is an ideal measure of sustainability; reserves to GDP in particular is
likely to reflect liquidity rather than solvency.12 We present specifications with and
without these variables to check that our results are robust to their inclusion and to
make our results more readily comparable to earlier studies.

3. Spreads and Macroeconomic Fundamentals

We now explore how much of the variation in sovereign yield spreads can be
explained by macroeconomic fundamentals by running linear regressions of yield
spreads on explanatory variables. For each country we construct an annual end-of-
year spread series to match the annual frequency of our measures of fundamentals.
Our measure of the end-of-year spread is the median daily EMBI spread from J.P.
Morgan for the month of December.13 Since we want to analyze determinants of

12
We thank an anonymous referee for helpful comments on these variables.
13
An alternative would be to use the daily spread on the last trading day of the year which would be
a more noisy measure. Our results are robust to using this alternative measure.
DETERMINANTS OF SOVEREIGN RISK 245

Table II. Summary statistics for EMBI spread regression sample

This table reports summary statistics for the spread regression sample (1994–2007). EMBI spread
refers to spreads on J.P. Morgan’s Emerging Market Bond Index Global. Volatility of terms of trade
is the standard deviation of the annual percentage change in terms of trade. It is calculated using
annual data from a ten year rolling window. Change in terms of trade is the percentage change in
terms of trade over the previous five years. Years since last default counts the number of years since
the last year in which the country was in default. It is capped at 10, equal to zero if the country is
in default, and set equal to 11 if the country has never defaulted. It is constructed using a default
indicator variable from Reinhart et al. (2003). Debt/GDP denotes U.S. dollar denominated debt to

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

GDP; reserves/GDP is the ratio of reserves (incluing gold) to GDP. The sample corresponds to the
regression sample used in Table III. p5 and p95 refer to the 5th and 95th percentiles of the distribution
respectively.

EMBI Volatility of Change in terms Years since

spread terms of trade of trade last default Debt/GDP Reserves/GDP

Mean 376 5.7 3.6 8.6 46.6 18.2

Median 265 4.4 0.2 10.0 44.5 14.8
St. Dev. 328 4.5 21.2 3.4 21.8 13.8
Min 24 0.7 −31.9 1 11.8 1.5
p5 60 1.5 −17.0 1 15.8 5.2
p95 1072 13.6 55.3 11 87.0 45.4
Max 1803 25.5 162.4 11 114.2 84.5
Number of observations: 276

the risk of sovereign default and its impact on debt prices, we focus on spread
observations while the country is not in default.
For an observation to be included in the regression we require that the full set of
explanatory variables is available. This leaves us with a regression sample consist-
ing of 276 country-year observations covering 31 countries from 1994 to 2007. We
restrict the sample to be the same for different specifications to facilitate compar-
ison of point estimates, significance levels, and adjusted R2 across specifications.
Table II reports summary statistics for the spread regression sample. It is apparent
that there is substantial variation in spreads and all of our explanatory variables.
Table III reports results from regressions of spreads on different sets of explana-
tory variables. We group the explanatory variables into three categories: our main
country-specific variables, global time series variables, and control variables. In col-
umn (1), we run a baseline regression with country-specific and global variables.
Volatility of terms of trade and changes in the terms of trade have the expected sign
and are significant at the 1% level: countries with higher terms of trade volatilities
and countries which have experienced a deterioration in their terms of trade tend
to have higher spreads. The coefficient on the years since last default variable is
negative and significant at the 1% level: the closer a country is to its most recent
default episode, the higher its spread tends to be. This latter result is consistent
Table III. Regressions of sovereign spreads on fundamentals

246
This table reports results from regressions of end-of-year (median December) EMBI spreads (31 countries, 1994–2007) on a set of explanatory
variables: macroeconomic variables, time series variables, control variables, and time and regional dummy variables. We only include observations
for which a country is not in default. In specifications (2) and (4) we use the volatility and change in the commodity price index to instrument for
the volatility and change of terms of trade. Credit ratings are from S&P; we include dummy variables corresponding to letter ratings in columns (8)
to (10). t-statistics (reported in parentheses) are calculated using standard errors which are robust and clustered by year. ∗∗ denotes significant at
1%; ∗ significant at 5%; + significant at 10%.
Regression equation:
spread = α + β1 vol tot + β2 chg tot + β3 ytd + β4 VIX + β5 DEF + β6 r 10year + β7 TED + β8 GDP debt + β reserves + dummies + ε
9
GDP
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

COUNTRY VARIABLES
Volatility of terms of trade (vol tot) 37.18 47.78 36.61 47.02 32.36 33.64 20.21 21.93
(9.79)∗∗ (7.76)∗∗ (9.90)∗∗ (8.65)∗∗ (8.75)∗∗ (8.81)∗∗ (2.76)∗ (3.25)∗∗
Change in terms of trade (chg tot) −5.04 −6.26 −4.33 −5.54 −3.86 −3.81 −2.43 −2.28
(5.54)∗∗ (4.63)∗∗ (5.44)∗∗ (4.52)∗∗ (4.98)∗∗ (4.88)∗∗ (2.21)∗ (2.22)∗
Years since last default (ytd) −35.22 −29.62 −25.19 −19.79 −29.34 −17.73 −23.20 −17.95
(5.71)∗∗ (4.32)∗∗ (3.43)∗∗ (2.50)∗ (3.87)∗∗ (3.43)∗∗ (3.65)∗∗ (2.46)∗
GLOBAL VARIABLES
VIX index (VIX) 9.33 9.92 7.66 8.27 7.46 15.24 15.07 11.49 10.31
(3.12)∗∗ (3.08)∗∗ (2.31)∗ (2.37)∗ (2.12)+ (3.67)∗∗ (3.63)∗∗ (3.52)∗∗ (3.08)∗∗
Default yield spread (DEF) 0.73 0.58 1.05 0.90 1.07 1.57 1.38 0.95 1.06

JENS HILSCHER AND YVES NOSBUSCH

(1.41) (1.20) (1.89)+ (1.69) (1.90)+ (1.65) (1.51) (1.43) (1.65)
Treasury 10-year yield (r 10year) 0.09 0.11 0.16 0.18 0.14 0.68 0.81 0.38 0.39
(0.42) (0.59) (0.76) (0.95) (0.67) (1.79)+ (2.23)∗ (1.41) (1.51)
TED spread (TED) 0.20 0.17 0.66 0.63 0.54 −0.74 −0.81 −0.26 0.07
(0.33) (0.29) (1.09) (1.04) (0.87) (1.11) (1.20) (0.42) (0.12)
CONTROL VARIABLES
Debt/GDP 4.16 4.09 4.13 4.25 2.50
(4.71)∗∗ (4.14)∗∗ (4.74)∗∗ (4.74)∗∗ (3.65)∗∗
Reserves/GDP −4.60 −4.59 −3.35 −2.66 −3.66
(3.56)∗∗ (3.39)∗∗ (2.83)∗ (2.19)∗ (3.11)∗∗
Credit rating X X X
Instrument for terms of trade X X
Regional effects X X
Year effects X
Number of observations 276 276 276 276 276 276 269 269 269 269
Adjusted R2 52.8% 51.5% 58.6% 57.3% 61.6% 66.9% 13.2% 52.3% 62.7% 65.1%

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

DETERMINANTS OF SOVEREIGN RISK 247

with the findings of Reinhart et al. (2003).14 The coefficient on the VIX index is
positive and significant at the 1% level; the coefficients on the other time series vari-
ables we include are not significant. The regression of spreads on country-specific
fundamentals and global factors delivers an adjusted R2 of 0.53. In contrast, the
regression in column (7) which only includes global factors, delivers a substantially
lower adjusted R2 of 0.13. We return to a discussion of global factors in the next
subsection.
We draw two main conclusions from these results. First, country-specific funda-

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

mentals measured by a country’s terms of trade have substantial explanatory power
for emerging market sovereign spreads, even after controlling for global factors.
Second, the volatility of these fundamentals is important in addition to their level.
Moreover the finding that a country-specific measure of terms of trade and the
associated volatility go a long way towards explaining emerging market sovereign
spreads is plausible on economic grounds given that most of these countries are
major commodity exporters.
A natural concern with this first regression is that the terms of trade could be
partly endogenous (Chen and Rogoff, 2003). In order to investigate this possibility,
we use an instrumental variables approach, instrumenting the terms of trade with our
commodity price indices. The idea is to use only major commodity prices traded
in world markets and to weight these prices by country-specific export shares.
While there may be legitimate concerns about endogeneity for other components
of terms of trade, the component related to commodities traded in world markets
is plausibly exogenous (Chen and Rogoff, 2003). The results for the instrumental
variable regression are reported in column (2). Change in terms of trade, volatil-
ity of terms of trade, and years since last default are still significant at the 1%
level. The instrumentation leads to a very modest drop in R2 of around 0.01. The
changes in coefficients on these variables are not significant at the 95% confidence
level. We interpret the results in column (2) as well as the instrumental variables
regression in column (4) for the slightly richer specification discussed in the next
paragraph to suggest that endogeneity of terms of trade is not a major issue for our
analysis.
Next, we include debt to GDP and reserves to GDP as additional control vari-
ables in columns (3) and (4). Many studies have found that debt to GDP has a
significant positive coefficient in regressions of spreads on levels of fundamen-
tals (e.g. Edwards, 1986; Eichengreen and Mody, 1998; Min, 1998). We confirm
this finding here as well as an intuitive negative coefficient on reserves to GDP.

14
An important question which we do not address in this paper is what factors aside from macroeco-
nomic fundamentals may cause some countries to be “serial defaulters.” The evidence in Kohlscheen
(2007) suggests that constitutions are important, in particular whether a country is a parliamentary
or presidential democracy.
248 JENS HILSCHER AND YVES NOSBUSCH

Both variables are significant at the 1% confidence level and including them leads
to an increase in the adjusted R2 of around 0.06. However, as noted in Section
2.3, these variables may be endogenous and should therefore be interpreted with
caution.
In column (5) of Table III, we investigate regional effects by including regional
dummy variables corresponding to the regions in Table I. We find that spread
levels are significantly higher in Latin America than in the other regions. Including
regional dummies increases the adjusted R2 by around 0.03; it does not lead to a

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

significant change in any of the coefficients.
In addition to being statistically significant, we find that our explanatory vari-
ables are economically significant. We multiply the coefficients from our main
specification in column (3) with each variable’s in-sample standard deviation. We
find that a one standard deviation increase in the volatility of terms of trade is
associated with an increase of 164 basis points in the spread. This magnitude un-
derlines the economic importance of the volatility of macroeconomic fundamentals
for sovereign spreads. The corresponding effects (in basis points) for the other
statistically significant explanatory variables are −92 for change in terms of trade,
−85 for years since last default, 45 for the VIX index, 90 for debt to GDP and −63
for reserves to GDP.
In summary, spreads are high for country-year observations where volatility of
terms of trade is high, terms of trade have recently deteriorated, the country has
recently defaulted, the VIX index is high, the level of debt to GDP is high, or
reserves are low.

3.1 GLOBAL FACTORS

We next examine the global time series variables (VIX index, U.S. default yield
spread, 10-year U.S. Treasury yield, TED spread) in more detail. Our first main
finding is that global factors are indeed important. In column (7) of Table III, we
report results from a regression that only includes our four time series variables.
Overall, these four variables alone give an adjusted R2 of 0.13. González-Rozada
and Levy Yeyati (2008) show that at higher (monthly or weekly) frequencies, the
explanatory power of the U.S. high yield corporate spread and Treasury rate is
even higher. Other papers demonstrating the importance of global factors include
Eichengreen and Mody (1998), Herrera and Perry (2002), Grandes (2003), Diaz
Weigel and Gemmill (2006) and Garcı́a-Herrero and Ortiz (2006).
The finding that the coefficient on the VIX index is positive and significant at
the 1% level is consistent with two recent studies using higher frequency emerging
market CDS spread data: Longstaff et al. (2007) find that at higher frequencies
there is evidence of comovement in sovereign CDS spreads while Pan and Singleton
DETERMINANTS OF SOVEREIGN RISK 249

(2008) find that the VIX is statistically significant in explaining CDS spreads of
Mexico, Turkey, and Korea.15
The finding of a positive coefficient on the 10-year U.S. Treasury yield is con-
sistent with the findings of Arora and Cerisola (2001) and Dailami et al. (2005),
among others. Other evidence on the effect of U.S. interest rates in this literature is
mixed. For instance, Kamin and von Kleist (1999) could not identify a robust rela-
tionship between a number of industrial country interest rates and emerging market
spreads. Min (1998) finds a positive but insignificant effect of U.S. interest rates on

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

sovereign credit spreads, while Eichengreen and Mody (2000) report a significant
negative effect. More recently, Uribe and Yue (2006) find that an increase in the
world interest rate causes a decline in emerging market spreads in the short run
followed by an overall increase in spreads in the long run.
More general contagion effects or sudden stops in international capital flows, as
in the work of Calvo et al. (1993) and many related papers (e.g. Calvo and Mendoza,
2000; Calvo, 2002; Calvo and Talvi, 2004; Calvo et al., 2006; González-Rozada
and Levy Yeyati, 2008) are also likely to be at work. Such effects may not be
fully captured by our time series variables. A more flexible way to capture these
effects is to include year dummies in the regression. When we replace time series
variables with year dummies in column (6), we indeed find that the coefficients on
the dummies for the crisis years 1995 and 1998 are positive and significant and
those for the four years from 2003 to 2006, a period marked by very low volatility
and low yields, are negative and significant. Moreover, comparing the R2 across
columns (5) and (6) shows that including year dummies instead of the time series
variables leads to an increase in adjusted R2 from 0.62 to 0.67. This suggests that
there is some comovement not fully captured by our four time series variables.
In addition to demonstrating the importance of global variables, the papers men-
tioned in this section typically find that local factors have little explanatory power
for emerging market sovereign spreads. While global factors are undeniably impor-
tant, comparing our results in columns (1) and (7) suggests that macroeconomic
fundamentals do matter. There are at least three possible reasons for this differ-
ence in findings. First, most papers focus on macroeconomic fundamentals other
than the terms of trade emphasized by our study. Second, they typically do not
include measures of the volatility of these fundamentals, whereas we find that
the volatility of terms of trade has substantial explanatory power. The focus on
other macroeconomic fundamentals of these studies may be partly driven by data
availability. Terms of trade are not available at the monthly or weekly frequencies
studied by González-Rozada and Levy Yeyati (2008) and Longstaff et al. (2007),
for example. Third, it may be that global factors are particularly important at high

15
In related work, Berndt et al. (2005) find that the VIX has substantial explanatory power in a high
frequency data set of U.S. corporate CDS spreads.
250 JENS HILSCHER AND YVES NOSBUSCH

frequencies, perhaps due to short run changes in aggregate liquidity or uncertainty,

while at low frequencies the explanatory power of country-specific macroeconomic
fundamentals and their volatility becomes relatively more important.

3.2 CREDIT RATINGS

A number of papers in the literature on emerging market sovereign debt focus

on credit ratings. One approach is to take credit ratings as a summary measure of

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

country-specific fundamentals (e.g. Kamin and von Kleist, 1999; González-Rozada
and Levy Yeyati, 2008; Powell and Martinez, 2008).16 A natural question is whether
our measures of country-specific fundamentals add explanatory power relative to
credit ratings.17 In order to investigate this, we include ratings dummies constructed
using S&P sovereign credit ratings. Comparing the results in columns (8) and (9)
of Table III we see that, compared to a regression that just includes time series
variables and credit ratings, adding volatility of terms of trade, change in the terms
of trade, and years since last default increases the adjusted R2 from 0.52 to 0.63.
In addition, the coefficients on our country-specific variables remain statistically
significant when we include ratings. Including debt to GDP and reserves to GDP
in column (10) leads to a further increase in adjusted R2 to 0.65. This increase in
explanatory power, and particularly the 0.11 increase in adjusted R2 from column
(8) to column (9), suggests that there is significant additional information in our
measures of country-specific fundamentals relative to credit ratings. One potential
explanation is that credit rating agencies do not fully take into account the risk
associated with terms of trade. This would be consistent with the findings of
Cantor and Packer (1996), Afonso et al. (2006), and Powell and Martinez (2008),
who study the determinants of sovereign credit ratings.18

3.3 ROBUSTNESS

We perform several robustness checks. First, we drop large oil exporting countries
from the regression. In particular we drop countries with more than 20% of exports
concentrated in crude petroleum and petroleum products. The results are largely
unaffected indicating that fluctuations in oil prices are only part of the reason why
terms of trade fluctuations impact sovereign spreads. In order to make sure that out-
liers are not driving some of our results we re-estimate the reduced form regressions

16
One issue in this context highlighted by González-Rozada and Levy Yeyati (2008) is the potential
endogeneity of credit ratings.
17
The question we are interested in here is to what extent credit ratings explain market prices.
A different question is whether credit ratings contain information about future fundamentals not
captured by market prices. Cavallo et al. (2008) find that they do.
18
In addition, Mora (2006) provides evidence of inertia in ratings.
DETERMINANTS OF SOVEREIGN RISK 251

on winsorized data where we winsorize the explanatory variables at the 5th and 95th
percentile levels, replacing values above the 95th percentile of the distribution with
the 95th percentile and values below the 5th percentile with the 5th percentile of the
distribution. Our main results are unaffected by these changes. Third, we replace the
volatility of terms of trade with its one year lag. We also drop observations in years
preceding a default. Our results are robust to these changes. Finally, we run our
main specification (3) in Table III in first differences as an additional specification
check. We find that the magnitudes of the coefficients on country fundamentals are

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

comparable.19

4. Default Probabilities and Spreads

In this section we estimate default probabilities directly in a hazard model. We use

fitted default probabilities to calculate predicted spreads and compare predicted
spreads to observed spreads out-of-sample.
We begin by constructing a measure of default which is equal to one in a given
year if the country goes into default over the course of the year (summarized in
Table I). Country defaults are rare events. In order to estimate a hazard model, we
need a large enough sample with a sufficient number of defaults. We take advantage
of the fact that default data are available for a longer period than spread data and
extend the sample back to 1970.20
We use the same set of explanatory variables as in the reduced form spread re-
gressions measured over the extended sample period of 1970 to 2007. We winsorize
the data to ensure that outliers are not driving our results. Conditional on availability
of covariates, the sample has 695 country-year observations, including 28 defaults
(a little less than 4% of the sample). Table IV presents summary statistics for our
explanatory variables. We group observations into two groups: those where the
country goes into default over the next year and those observations for which it
stays solvent. We calculate statistics for non-default and default observations sep-
arately. For each of the explanatory variables the mean and median of the default
observations reflect less favorable conditions. Volatility of terms of trade tends to
be particularly unfavorable in country-years preceding default: both the mean and
median of the default observation sample are about one standard deviation larger
than the mean and median for the non-default sample.
19
The coefficients on the default yield spread and the TED spread remain positive. Consistent
with Uribe and Yue (2006), we find that changes in U.S. interest rates are negatively related to
contemporaneous changes in spreads (but positively related to future changes in spreads). VIX loses
statistical significance.
20
Although default data extend back farther, terms of trade data are available only from 1960. Since
we calculate volatility over a ten year backward looking window the sample with available covariates
begins in 1970.
252 JENS HILSCHER AND YVES NOSBUSCH

Table IV. Summary statistics for default prediction

This table reports summary statistics for our set of country variables and control variables for the
default prediction sample (1970–2007). We consider the sample for which all variables are available.
Variables are winsorized at the 5th and 95th percentile levels. In Panel A we report summary statistics
of the group of observations for which the country does not enter default over the following year and
in Panel B we report statistics for those observations when the country does enter default over the
following year. The sample corresponds to the sample used for logit regressions in Table V.

Volatility of Change in Years since

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

terms of trade terms of trade last default Debt/GDP Reserves/GDP

Panel A: Non-default observations (not in default next period)

Mean 8.3 0.4 9.5 41.5 12.8
Median 7.3 −0.3 11 39.1 10.3
St. Dev. 5.5 15.9 2.9 19.6 8.4
Min 2.1 −31.6 1 12.4 1.6
Max 23.3 39.1 11 103.3 31.6
Number of observations: 695
Panel B: Default observations (in default next period)
Mean 13.7 −1.1 7.5 57.9 9.3
Median 13.6 −3.5 11 57.4 5.5
St. Dev. 6.3 21.1 4.4 24.7 7.9
Number of observations: 28

In order to investigate the predictive power of the explanatory variables more

formally, we estimate a logit model relating a forward-looking default indicator to
explanatory variables. Estimation results for different specifications are reported in
Table V. We first consider a specification that includes volatility of terms of trade,
the change in the terms of trade, and years since last default. Volatility of terms of
trade and years since last default are significant at the 1% level. Next we include
debt to GDP and reserves to GDP, which both enter with the expected sign and
are significant at the 1% and 5% levels respectively. Volatility of terms of trade
and years since last default remain significant at the 1% and 10% levels. Model fit
measured by pseudo R2 improves from 0.13 to 0.20.21 Following our approach for
estimating EMBI spread regressions, we also include two specifications in which
we instrument for terms of trade. We find that model fit, coefficient magnitudes,
and coefficient significance levels are comparable, except in the case of years
since last default, for which the estimated coefficient decreases in magnitude and
loses statistical significance. We refer to the model in column (3) as our main
specification. It includes the country variables and control variables that correspond

21
These results are consistent with independent work by Catão and Sutton (2002) and Catão and
Kapur (2006) who also investigate empirical determinants of default in a reduced form hazard
model setting that is similar to the specification reported in Table V. These authors do not, however,
instrument for terms of trade.
DETERMINANTS OF SOVEREIGN RISK 253

Table V. Reduced form default prediction

This table reports results from logit regressions of a forward looking default indicator on explanatory
variables (1970–2007). The default indicator Yi,t+1 is equal to one if country i goes into default during
the next year (t + 1) and zero otherwise. Summary statistics for the sample are reported in Table IV.
z-statistics are reported in parentheses. ∗∗ denotes significant at 1%; ∗ significant at 5%; + significant
at 10%.
Regression equation:
  
P (Yi,t+1 = 1) = 1 + exp −α − β1 vol tot − β2 chg tot − β3 ytd − β4 GDP debt − β reserves − ε −1
5
GDP

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

(1) (2) (3) (4)

Volatility of terms of trade (vol_tot) 0.141 0.138 0.140 0.114

(4.61)∗∗ (2.67)∗∗ (4.39)∗∗ (1.77)+
Change in terms of trade (chg_tot) −0.006 −0.005 0.010 0.002
(0.63) (0.13) (0.86) (0.04)
Years since last default (ytd) −0.155 −0.065 −0.102 −0.017
(3.19)∗∗ (1.12) (1.90)+ (0.28)
Debt/GDP 0.038 0.033
(3.73)∗∗ (2.95)∗∗
Reserves/GDP −0.068 −0.102
(2.14)∗ (2.18)∗
Constant −3.40 −3.69 −4.95 −4.45
(5.93)∗∗ (4.86)∗∗ (5.04)∗∗ (4.36)∗∗

Instrument for terms of trade X X

Number of observations 723 541 723 541

Number of defaults 28 24 28 24

Pseudo R2 12.8% 11.6% 20.5% 21.7%

Accuracy ratio 76.7% 77.5% 83.3% 84.7%

to the main specification in Table III.22 To get a sense of economic significance we

compute the proportional change of the predicted default probability comparing the
case when all explanatory variables are equal to their mean values to the case where
we set each variable individually equal to one sample standard deviation above its
mean. This leads to an increase in the default probability of 114% for volatility of
terms of trade and 110% for debt to GDP and a decrease in default probability of
26% for years since last default and 43% for reserves to GDP.

22
In principle, we could consider additional explanatory variables that predict default in the sample
we consider. However, because of the sample size and the rather small number of default observations,
we choose only the country variables and control variables included in our main spread regression
specification.
254 JENS HILSCHER AND YVES NOSBUSCH

2500

1000

500

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

100

50

10

10 50 100 500 1000 2500

Predicted_spread

EMBI_spread Fitted_values

Figure 3. Actual and predicted spreads

This graph plots EMBI spreads (from J.P. Morgan, end-of-year spread over U.S. Treasuries) against
out-of-sample predicted spreads using the logit model from Table V to calculate fitted probabilities
of default.

4.1 OUT-OF-SAMPLE COMPARISON OF PREDICTED AND EMBI SPREADS

We next calculate predicted spreads using fitted default probabilities from the hazard
model and compare these to observed EMBI spreads. We assume that recovery rates
are fixed, i.e. that creditors receive a constant fraction of face value in the event
of default. We set the fractional recovery rate equal to 0.6 which is broadly in line
with the average historical experience (Sturzenegger and Zettelmeyer, 2005).23 We
compute the default component of the spread by discounting the expected payoff at
the riskfree rate and calculate predicted spreads using fitted probabilities from our
main specification in column (3) of Table V. We estimate the model using data up
to 1993 and then calculate fitted default probabilities for the period 1994 to 2007.
We compare out-of-sample predicted spreads to actual spreads and ask how much
of the level and variation in EMBI spreads is due to country default probabilities.
Figure 3 plots actual against predicted spreads. We notice a strong positive relation
which is reflected in a correlation of 0.65.

23
Panizza et al. (2008) provide a comprehensive discussion of the legal aspects of sovereign debt
and default.
DETERMINANTS OF SOVEREIGN RISK 255

Table VI. Comparing actual and out-of-sample predicted spreads

This table reports results of comparing actual to predicted spreads. Panel A reports summary statistics
for EMBI spreads and predicted spreads. The sample corresponds to the spread regression sample
of Table III. Panel B reports estimates from regressions of EMBI spreads on predicted spreads. To
control for outliers, predicted spreads are winsorized, which results in replacing the two largest and the
two smallest spread observations. t-statistics (reported in parentheses) are calculated using standard
errors which are robust and clustered by year. ∗∗ denotes significant at 1%; ∗ significant at 5%;
+
significant at 10%.
Regression equation (Panel B):

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

spread = α + β1 spread pred + β2 VIX + β3 DEF + β4 r 10year + β5 TED + dummies + ε

Panel A: Summary statistics of predicted spreads and default probabilities

EMBI Predicted
spread spread

Mean 376 229

Median 265 87
St. Dev. 328 426
p5 60 17
p95 1072 941
Observations: 276

Panel B: Regression of actual on predicted spreads

(1) (2) (3) (4) (5)

Predicted spread (spread_pred) 0.72 0.67 0.41

(10.70)∗∗ (9.49)∗∗ (4.56)∗∗
GLOBAL VARIABLES
VIX index (VIX) 10.4 15.2 15.1 12.6
(2.70)∗ (3.67)∗∗ (3.63)∗∗ (3.39)∗∗
Default yield spread (DEF) 1.0 1.6 1.4 1.0
(1.45) (1.65) (1.51) (1.38)
Treasury 10-year yield (r_10year) 0.31 0.68 0.81 0.56
(1.32) (1.79)+ (2.23)∗ (2.21)∗
TED spread (TED) 0.13 −0.74 −0.81 −0.34
(0.20) (1.11) (1.20) (0.58)
CONTROL VARIABLES
Credit rating X X

Constant 222.9 −225.4 −379.4 −616.2 −505.6

(7.18)∗∗ (1.77)+ (1.82)+ (2.87)∗ (3.57)∗∗

Number of observations 276 276 269 269 269

Adjusted R2 44.8% 49.0% 13.2% 52.3% 61.3%

Table VI Panel A reports summary statistics for EMBI spreads and predicted
spreads. The sample contains spread observations from 31 countries over the period
1994 to 2007 and corresponds to the regression sample used in Table III. Predicted
spreads are smaller than actual spreads: the median predicted spread is equal to
256 JENS HILSCHER AND YVES NOSBUSCH

87 basis points, compared to a median EMBI spread of 265 basis points. This
difference reflects non-default spread components, to which we return later in this
section.
We check more formally how much of the variation in spreads is explained by
variation in default risk by regressing EMBI spreads on predicted spreads and
global variables. Predicted spreads have some extreme values. In order to control
for outliers we winsorize predicted spreads. As a result of this winsorization the
in-sample standard deviation is reduced from 426 to 307 and the kurtosis is reduced

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

from 64 to 9. We report regression results in Table VI Panel B. We first include
only predicted spreads. The regression coefficient is 0.72, which is statistically
significant at the 1% level; the R2 is 0.45. This is a surprisingly high level of
explanatory power, keeping in mind that our model estimation does not use as
inputs any of the observed spreads it is attempting to explain. For comparison, our
main specification in Table III explains 62% of the variation.
We next investigate the effect of adding global variables and credit ratings. When
we add the four time series variables (column (2)), the VIX index enters with a
positive coefficient and is statistically significant while the other time series vari-
ables are not significant. Adding global variables increases the adjusted R2 by
0.04. However, we again find that predicted spreads contain important informa-
tion not captured by global variables: including only global variables delivers a
much lower adjusted R2 of 0.13 (column (3)). Predicted spreads also contain in-
formation relative to credit ratings. Predicted spreads enter significantly when we
include them together with credit ratings (column (5)) and increase the adjusted R2
by 0.09.
We explore if there is variation across different levels of credit risk in the fraction
of the spread that is explained by default risk. We group spreads by their esti-
mated default probability and compare average observed and fitted spreads across
groups. We choose break points using the distribution of fitted probabilities from
the regression sample (1994 to 2007). Figure 4 plots average EMBI spreads and
predicted spreads by default probability quintiles. We find that fitted spreads for
good credit are much smaller than observed spreads while for poor credit predicted
and observed spreads are more similar in magnitude. These findings are consistent
with patterns in the corporate bond market. Huang and Huang (2003) show that
spreads implied by structural form models are far too low on companies of good
credit quality but they are closer to actual spreads for companies of poor credit
quality.
Comparing means of observed and model implied spreads in Table VI Panel A or
in Figure 4 shows that a large component of the spread is not explained by our model.
This finding is not surprising given that we only estimate the default component
of the spread. It is related to several studies that investigate non default-risk spread
components in the sovereign and corporate bond markets. Hund and Lesmond
DETERMINANTS OF SOVEREIGN RISK 257

800
Average spread (basis points)

600

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

400

200

0
.6 1.1 2.1 4.5 11.9

EMBI spread Predicted spread

Figure 4. Spreads by average default probability

This graph plots average EMBI spreads (from J.P. Morgan, end-of-year spread over U.S. Treasuries)
and predicted spreads by fitted default probability quintiles. We report the average default probability
for each quintile on the horizontal axis. The sample used corresponds to that in Table VI.

(2007) provide evidence that a significant part of the spread on emerging market
corporate and sovereign debt may be explained by liquidity. Borri and Verdelhan
(2008), Martell (2008), Remolana et al. (2008), and Andrade (2009) also investigate
non-default risk determinants of spreads in the sovereign bond market.24
To summarize, we find that predicted spreads have significant explanatory power
for observed sovereign spreads. The default component of spreads explains variation
in observed spreads surprisingly well and the fraction of the spread due to default
risk is much larger for borrowers of lower credit quality.

24
For risky corporate bonds, Huang and Huang (2003) find that default risk accounts for only part
of the spread. Elton et al. (2001) document important spread components related to liquidity, risk
aversion, and taxes. Chen et al. (2007) present evidence that liquidity is priced in U.S. corporate
bonds. Longstaff et al. (2005) find that the majority of the U.S. corporate default swap spread is
explained by default risk and that the non-default component is time-varying and closely related to
measures of liquidity.
258 JENS HILSCHER AND YVES NOSBUSCH

5. Conclusion

In this paper we provide evidence that variation in country fundamentals explains

a large share of the variation in emerging market sovereign debt prices for a set
of 31 countries over the period 1994 to 2007. In a departure from the existing
empirical literature on sovereign debt we pay special attention to the volatility of
fundamentals. In particular we find that the volatility of terms of trade is both
statistically and economically significant in explaining spread variation. A one

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

standard deviation increase in the volatility of terms of trade is associated with an
increase of 164 basis points in spreads, which corresponds to around half of the
standard deviation of observed spreads. The explanatory power of changes in terms
of trade and volatility of terms of trade is substantial, even after controlling for
global factors and sovereign credit ratings. It is robust to instrumenting for terms
of trade with commodity price indices.
We also analyze the empirical link between default probabilities and sovereign
debt prices. We estimate a hazard model using historical data on measures of
fundamentals and default and we find that the volatility of terms of trade is also
an important predictor of country default. We calculate spreads implied by our
default prediction model and find that the model implied default component of
spreads explains a large share of the variation in observed spreads out-of-sample.
The model fits better for borrowers of lower credit quality.
The findings in this paper have implications for investors and policymakers.
We have identified factors explaining variation in spreads across countries and
over time. Countries with larger fluctuations in their terms of trade, possibly due
to concentrated exports in commodities with volatile prices, are more susceptible
to external shocks. They may find it beneficial to address such risk factors in order to
reduce borrowing costs as well as the probability of a crisis or default. In particular,
countries could try to explicitly hedge exposures to macroeconomic shocks in
international financial markets. Caballero (2003), Shiller (2003), Merton (2005),
and a recent report by the Inter-American Development Bank (IDB report, 2006)
advocate the use of innovative financial contracts to share these macroeconomic
risks more effectively.

Appendix

J.P. Morgan’s EMBI includes U.S. dollar denominated Brady bonds, loans, and
Eurobonds issued or guaranteed by emerging market governments. Countries need
to be classified as low or middle income by the World Bank for the last two years,
have restructured debt in the past 10 years, or have restructured debt outstanding.
The minimum issue size for a debt instrument is U.S. dollar 500 million, it needs to
DETERMINANTS OF SOVEREIGN RISK 259

have a maturity greater than 2.5 years, verifiable daily prices and cash flows, and it
has to fall under G7 legal jurisdiction.25
The Commodity Research Bureau (CRB) commodity index series is from Global
Financial Data. According to the CRB, the index includes energy (crude oil, heating
oil, natural gas), grains and oilseed (corn, soybeans, wheat), industrials (copper,
cotton), livestock (live cattle, live hogs), precious metals (gold, platinum, silver),
softs (cocoa, coffee, orange juice, sugar).26
We use annual terms of trade data provided by the Development Data Group at the

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

World Bank. Terms of trade indices are constructed as the ratio of an export price in-
dex to an import price index. The underlying price and volume indices are compiled
by the United Nations Conference on Trade and Development (UNCTAD).
Our measure of government external debt is from the World Bank Global De-
velopment Finance data set. GDP data are from the International Monetary Fund
World Economic Outlook. Reserves refers to the Total Reserves Including Gold
series from the World Development Indicators.
Our commodity price index is constructed using data on export shares from
COMTRADE and commodity prices from Global Financial Data. The SITC codes
corresponding to the 12 components of our index are oil (33), coffee (071), textiles
(65), copper (682), cotton (263), cocoa (072), meat (011), bananas (0573), leather
(61), gold (97), gas (34), silver (681). The commodity price change is calculated
as the weighted average price change of the 12 price series included in the index,
where the weights are the annual export shares for the country.
The 10-year U.S. Treasury yield and the 3-month Treasury Bill rate are from
the Federal Reserve Bank of St. Louis. 3-month Libor is from Global Insight.
Yields on Aaa and Baa U.S. corporate bonds are from Global Financial Data. The
VIX measure of volatility of the S&P500 index is provided by the Chicago Board
Options Exchange.

References

Afonso, A., Gomes, P., and Rother, P. (2006) What “Hides” Behind Sovereign Debt Ratings? unpub-
lished working paper, European Central Bank.
Anderson, R. and Sundaresan, S. (2000) A Comparative Study of Structural Models of Corporate
Bond Yields: An Exploratory Investigation, Journal of Banking and Finance 24, 255–269.
Andrade, S. (2009) A Model of Asset Pricing Under Country Risk, Journal of International Money
and Finance 28, 671–695.
Arora, V. B. and Cerisola, M. B. (2001) How Does U.S. Monetary Policy Influence Economic
Conditions in Emerging Markets? IMF Staff Papers 48, 474–498.
Beck, R. (2001) Do Country Fundamentals Explain Emerging Market Bond Spreads? CFS Working
Paper, No. 2001/02, Johann Wolfgang Goethe-Universität.
25
Source: Emerging Markets External Debt Handbook for 2002–2003, J.P. Morgan.
26
Source: documentation describing CRB Commodity Index in Global Financial Data.
260 JENS HILSCHER AND YVES NOSBUSCH

Berg, A., Borensztein, E., Milesi-Ferretti, G. M., and Pattillo, C. (2000) Anticipating Balance of
Payments Crises–The Role of Early Warning Systems, IMF Occasional Papers 186, International
Monetary Fund.
Berg, A., Borensztein, E., and Pattillo, C. (2005) Assessing Early Warning Systems: How Have They
Worked in Practice? IMF Staff Papers 52, 462–502.
Berndt, A., Douglas, R., Duffie, D., Ferguson, M., and Schranz, D. (2005) Measuring
Default-Risk Premia from Default Swap Rates and EDFs, unpublished working paper, Tep-
per School of Business, Carnegie Mellon University, and Stanford Graduate School of
Business.

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

Borri, N. and Verdelhan, A. (2008) Sovereign Risk Premia, unpublished working paper, Boston
University.
Bulow, J. and Rogoff, K. (1989) Sovereign Debt: Is to Forgive to Forget? American Economic Review
79, 43–50.
Caballero, R. J. (2003) The Future of the IMF, American Economic Review Papers and Proceedings
93, 31–38.
Calvo, G. A. (2002) Globalization Hazard and Delayed Reforms in Emerging Markets, Economia 2,
1–29.
Calvo, G. A., Izquierdo, A., and Talvi, E. (2006) Sudden Stops and Phoenix Miracles in Emerging
Markets, American Economic Review 96, 405–410.
Calvo, G. A., Leiderman, L., and Reinhart, C. M. (1993) Capital Inflows and Real Exchange Rate
Appreciation in Latin America, IMF Staff Papers 40, 108–151.
Calvo, G. A., and Mendoza, E. G. (1996) Mexico’s Balance-of-Payments Crisis: A Chronicle of a
Death Foretold, Journal of International Economics 41, 235–264.
Calvo, G. A. and Mendoza, E. G. (2000) Rational Contagion and the Globalization of Securities
Markets, Journal of International Economics 51, 79–113.
Calvo, G. A. and Talvi, E. (2004) Sudden Stops, Financial Factors and Economic Collapse in Latin
America: Learning from Argentina and Chile, NBER Working Paper 11153.
Campbell, J. Y. and Taksler, G. B. (2003) Equity Volatility and Corporate Bond Yields, Journal of
Finance 58, 2321–2349.
Cantor, R. and Packer, F. (1996) Determinants and Impact of Sovereign Credit Ratings, Federal
Reserve Bank of New York Economic Policy Review 37–54.
Catão, L. and Kapur, S. (2006) Volatility and the Debt Intolerance Paradox, IMF Staff Papers 53,
195–218.
Catão, L. and Sutton, B. (2002) Sovereign Defaults: The Role of Volatility, IMF Working Paper,
No. 02/149, International Monetary Fund.
Cavallo, E. A., Powell, A., and Rigobon, R. (2008) Do Credit Rating Agencies Add Value? Evi-
dence from the Sovereign Rating Business Institutions, unpublished working paper, Inter-American
Development Bank and Massachusetts Institute of Technology.
Chen, L., Lesmond, D. A., and Wei, J. (2007) Corporate Yield Spreads and Bond Liquidity, Journal
of Finance 62, 119–149.
Chen, Y.-C. and Rogoff, K. (2003) Commodity Currencies, Journal of International Economics 60,
133–160.
Dailami, M., Masson, P., and Padou, J. J. (2005) Global Monetary Conditions versus Country-Specific
Factors in the Determination of Emerging Market Debt Spreads, World Bank Policy Research Paper
3626.
Diaz Weigel, D. and Gemmill, G. (2006) What Drives Credit Risk in Emerging Markets? The Roles
of Country Fundamentals and Market Co-Movements, Journal of International Money and Finance
25, 476–502.
DETERMINANTS OF SOVEREIGN RISK 261

Duffie, D., Pedersen, L. H., and Singleton, K. J. (2003) Modeling Sovereign Yield Spreads: A Case
Study of Russian Debt, Journal of Finance 58, 119–159.
Edwards, S. (1984) LDC Foreign Borrowing and Default Risk: An Empirical Investigation, 1976–80,
American Economic Review 74, 726–734.
Edwards, S. (1986) The Pricing of Bonds and Bank Loans in International Markets, An Empirical
Analysis of Developing Countries’ Foreign Borrowing, European Economic Review 30, 565–589.
Eichengreen, B. and Mody, A. (2000) What Explains Changing Spreads on Emerging-Market
Debt: Fundamentals or Market Sentiment?, in: S. Edwards (ed.), Capital Flows and the Emerg-
ing Economies: Theory, Evidence and Controversies, University of Chicago Press.

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

Elton, E. J., Gruber, M. J., Agrawal, D., and Mann, C. (2001) Explaining the Rate Spread on Corporate
Bonds, Journal of Finance 56, 247–277.
Emerging Market Bond Index, J. P. Morgan, New York, NY.
Emerging Markets External Debt Handbook for 2002–2003, J. P. Morgan, New York, NY.
Favero, C. and Giavazzi, F. (2002) Why are Brazil’s Interest Rates so High? unpublished working
paper, IGIER-Bocconi.
Favero, C. and Giavazzi, F. (2005) Inflation Targeting and Debt: Lessons from Brazil, in: F. Giavazzi,
I. Goldfajn and S. Herrera (eds.), Inflation Targeting, Debt and the Brazilian experience 1999 to
2003, MIT Press.
Ferrucci, G. (2003) Empirical Determinants of Emerging Market Economies’ Sovereign Bond
Spreads, Bank of England Working Paper 205.
Garcia, M. and Rigobon, R. (2005) A Risk Management Approach to Emerging Market’s Sovereign
Debt Sustainability with an Application to Brazilian Data. in: F. Giavazzi, I. Goldfajn, and S. Herrera
(eds.), Inflation Targeting, Debt and the Brazilian experience 1999 to 2003, MIT Press.
Garcı́a-Herrero, A. and Ortı́z, A. (2006) The Role of Global Risk Aversion in Explaining Sovereign
Spreads, Economia 7, 125–155.
Goldstein, M., Kaminsky, G. L., and Reinhart, C. M. (2000) Assessing Financial Vulnerability, An
Early Warning System for Emerging Markets, Institute for International Economics, Washington, DC.
González-Rozada, M., and Levy Yeyati, E. (2008) Global Factors and Emerging Market Spreads,
Economic Journal 118, 1917–1936.
Grandes, M. (2003) Convergence and Divergence in Sovereign Bond Spreads: Theory and Facts from
Latin America, unpublished working paper, Ecole Normale Supérieure, Paris.
Hajivassiliou, V. A. (1987) The External Debt Repayments Problems of LDC’s, An Econometric
Model Based on Panel Data, Journal of Econometrics 36, 205–230.
Hajivassiliou, V. A. (1994) A Simulation Estimation Analysis of the External Debt Crises of Devel-
oping Countries, Journal of Applied Econometrics 9, 109–131.
Herrera, S. and Perry, G. (2002) Determinants of Latin Spreads in the New Economy Era: The Role
of U.S. Interest Rates and Other External Variables, unpublished working paper, World Bank.
Huang, J.-Z. and Huang, M. (2003) How Much of the Corporate-Treasury Yield Spread is Due to
Credit Risk? unpublished working paper, Smeal College of Business, Pennsylvania State University
and Graduate School of Business, Stanford University.
Hund, J. and Lesmond, D. A. (2007) Liquidity and Credit Risk in Emerging Debt Markets, unpublished
working paper, University of Texas - Austin and Tulane University.
Global Development Finance, World Bank, Washington, DC, various issues.
IDB report (2006) Living with Debt: How to Limit the Risks of Sovereign Finance, Inter-American
Development Bank, Washington, DC.
Kamin, S. B. and von Kleist, K. (1999) The Evolution and Determinants of Emerging Market Credit
Spreads in the 1990’s, Board of Governors of the Federal Reserve System Internatinal Finance
Discussion Paper 653.
262 JENS HILSCHER AND YVES NOSBUSCH

Kaminsky, G. L. and Reinhart, C. M. (1999) The Twin Crises: The Causes of Banking and Balance-
of-Payments Problems, American Economic Review 89, 473–500.
Kohlscheen, E. (2007) Why are there Serial Defaulters? Evidence from Constitutions, The Journal
of Law and Economics 50, 713–30.
Longstaff, F. A., Mithal, S., and Neis, E. (2005) Corporate Yield Spreads: Default Risk or Liquidity?
New Evidence from the Credit Default Swap Market, Journal of Finance 60, 2213–2253.
Longstaff, F. A., Pan, J., Pedersen, L. H., and Singleton, K. J. (2007) How Sovereign is Sovereign
Credit Risk? unpublished working paper, UCLA Anderson School, MIT Sloan School, NYU Stern
School, and Stanford Graduate School of Business.

Downloaded from http://rof.oxfordjournals.org/ at Belgorod State University on December 8, 2013

Martell, R. (2008) Understanding Common Factors in Domestic and International Bond Spreads,
Review of Finance 12, 365–389.
Martinez, J. F. and Powell, A. (2008) On Emerging Market Sovereign Spreads and Ratings, unpub-
lished working paper, Inter-American Development Bank and Oxford University.
Mendoza, E. G. (1995) The Terms of Trade, The Real Exchange Rate and Economic Fluctuations,
International Economic Review 36, 101–137.
Mendoza, E. G. (1997) Terms of Trade Uncertainty and Economic Growth, Journal of Development
Economics 54, 323–356.
Merton, R. C. (1974) On the Pricing of Corporate Debt: The Risk Structure of Interest Rates, Journal
of Finance 29, 449–470.
Merton, R. C. (2005) Swapping Your Country’s Risks, Harvard Business Review 83, 34–36.
Min, H. G. (1998) Determinants of Emerging Market Bond Spreads: Do Economic Fundamentals
Matter? World Bank Policy Research Paper 1899.
Mora, N. (2006) Sovereign Credit Ratings: Guilty Beyond Reasonable Doubt? Journal of Banking
and Finance 30, 2041–2062.
Pan, J. and Singleton, K. J. (2008) Default and Recovery Implicit in the Term Structure of Sovereign
CDS Spreads, Journal of Finance 63, 2345–2384.
Panizza, U., Sturzenegger, F., and Zettelmeyer, J. (2008) The Economics and Law of Sovereign Debt
and Default, forthcoming, Journal of Economic Literature.
Reinhart, C. M., Rogoff, K. S., Savastano, M. A. (2003) Debt Intolerance, in: W. Brainard and G.
Perry (eds.), Brookings Papers on Economic Activity 1, 1–74.
Remolana, E., Scatigna, M., and Wu, E. (2008) The Dynamic Pricing of Sovereign Risk in Emerg-
ing Markets: Fundamentals and Risk Aversion, unpublished working paper, Bank for International
Settlements and University of New South Wales.
Shiller, R. J. (2003) The New Financial Order: Risks in the 21st Century, Princeton University Press,
Princeton, NJ.
Shumway, T. (2001) Forecasting Bankruptcy More Accurately: A Simple Hazard Model, Journal of
Business 74, 101–124.
Sturzenegger, F. and Zettelmeyer, J. (2005) Haircuts: Estimating Investor Losses in Sovereign Debt
Restructurings, 1998–2005, IMF Working Paper, No. 05/137, International Monetary Fund.
United Nations Conference on Trade and Development, United Nations, New York, NY, various
issues.
Uribe, M. and Yue, V. Z. (2006) Country Spreads and Emerging Countries: Who Drives Whom?
Journal of International Economics 69, 6–36.
Westphalen, M. (2001) Determinants of Sovereign Bond Credit Spreads Changes, unpublished work-
ing paper, Université de Lausanne.
World Economic Outlook, International Monetary Fund, Washington, DC.